src/du.c

   1 /* du -- summarize disk usage
   2    Copyright (C) 1988-1991, 1995-2011 Free Software Foundation, Inc.
   3
   4    This program is free software: you can redistribute it and/or modify
   5    it under the terms of the GNU General Public License as published by
   6    the Free Software Foundation, either version 3 of the License, or
   7    (at your option) any later version.
   8
   9    This program is distributed in the hope that it will be useful,
  10    but WITHOUT ANY WARRANTY; without even the implied warranty of
  11    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12    GNU General Public License for more details.
  13
  14    You should have received a copy of the GNU General Public License
  15    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
  16
  17 /* Differences from the Unix du:
  18    * Doesn't simply ignore the names of regular files given as arguments
  19      when -a is given.
  20
  21    By tege@sics.se, Torbjorn Granlund,
  22    and djm@ai.mit.edu, David MacKenzie.
  23    Variable blocks added by lm@sgi.com and eggert@twinsun.com.
  24    Rewritten to use nftw, then to use fts by Jim Meyering.  */
  25
  26 #include <config.h>
  27 #include <getopt.h>
  28 #include <sys/types.h>
  29 #include <assert.h>
  30 #include "system.h"
  31 #include "argmatch.h"
  32 #include "argv-iter.h"
  33 #include "di-set.h"
  34 #include "error.h"
  35 #include "exclude.h"
  36 #include "fprintftime.h"
  37 #include "human.h"
  38 #include "quote.h"
  39 #include "quotearg.h"
  40 #include "stat-size.h"
  41 #include "stat-time.h"
  42 #include "stdio--.h"
  43 #include "xfts.h"
  44 #include "xstrtol.h"
  45
  46 extern bool fts_debug;
  47
  48 /* The official name of this program (e.g., no `g' prefix).  */
  49 #define PROGRAM_NAME "du"
  50
  51 #define AUTHORS \
  52   proper_name_utf8 ("Torbjorn Granlund", "Torbj\303\266rn Granlund"), \
  53   proper_name ("David MacKenzie"), \
  54   proper_name ("Paul Eggert"), \
  55   proper_name ("Jim Meyering")
  56
  57 #if DU_DEBUG
  58 # define FTS_CROSS_CHECK(Fts) fts_cross_check (Fts)
  59 #else
  60 # define FTS_CROSS_CHECK(Fts)
  61 #endif
  62
  63 /* A set of dev/ino pairs.  */
  64 static struct di_set *di_set;
  65
  66 /* Keep track of the preceding "level" (depth in hierarchy)
  67    from one call of process_file to the next.  */
  68 static size_t prev_level;
  69
  70 /* Define a class for collecting directory information. */
  71 struct duinfo
  72 {
  73   /* Size of files in directory.  */
  74   uintmax_t size;
  75
  76   /* Latest time stamp found.  If tmax.tv_sec == TYPE_MINIMUM (time_t)
  77      && tmax.tv_nsec < 0, no time stamp has been found.  */
  78   struct timespec tmax;
  79 };
  80
  81 /* Initialize directory data.  */
  82 static inline void
  83 duinfo_init (struct duinfo *a)
  84 {
  85   a->size = 0;
  86   a->tmax.tv_sec = TYPE_MINIMUM (time_t);
  87   a->tmax.tv_nsec = -1;
  88 }
  89
  90 /* Set directory data.  */
  91 static inline void
  92 duinfo_set (struct duinfo *a, uintmax_t size, struct timespec tmax)
  93 {
  94   a->size = size;
  95   a->tmax = tmax;
  96 }
  97
  98 /* Accumulate directory data.  */
  99 static inline void
 100 duinfo_add (struct duinfo *a, struct duinfo const *b)
 101 {
 102   a->size += b->size;
 103   if (timespec_cmp (a->tmax, b->tmax) < 0)
 104     a->tmax = b->tmax;
 105 }
 106
 107 /* A structure for per-directory level information.  */
 108 struct dulevel
 109 {
 110   /* Entries in this directory.  */
 111   struct duinfo ent;
 112
 113   /* Total for subdirectories.  */
 114   struct duinfo subdir;
 115 };
 116
 117 /* If true, display counts for all files, not just directories.  */
 118 static bool opt_all = false;
 119
 120 /* If true, rather than using the disk usage of each file,
 121    use the apparent size (a la stat.st_size).  */
 122 static bool apparent_size = false;
 123
 124 /* If true, count each hard link of files with multiple links.  */
 125 static bool opt_count_all = false;
 126
 127 /* If true, hash all files to look for hard links.  */
 128 static bool hash_all;
 129
 130 /* If true, output the NUL byte instead of a newline at the end of each line. */
 131 static bool opt_nul_terminate_output = false;
 132
 133 /* If true, print a grand total at the end.  */
 134 static bool print_grand_total = false;
 135
 136 /* If nonzero, do not add sizes of subdirectories.  */
 137 static bool opt_separate_dirs = false;
 138
 139 /* Show the total for each directory (and file if --all) that is at
 140    most MAX_DEPTH levels down from the root of the hierarchy.  The root
 141    is at level 0, so `du --max-depth=0' is equivalent to `du -s'.  */
 142 static size_t max_depth = SIZE_MAX;
 143
 144 /* Human-readable options for output.  */
 145 static int human_output_opts;
 146
 147 /* If true, print most recently modified date, using the specified format.  */
 148 static bool opt_time = false;
 149
 150 /* Type of time to display. controlled by --time.  */
 151
 152 enum time_type
 153   {
 154     time_mtime,                 /* default */
 155     time_ctime,
 156     time_atime
 157   };
 158
 159 static enum time_type time_type = time_mtime;
 160
 161 /* User specified date / time style */
 162 static char const *time_style = NULL;
 163
 164 /* Format used to display date / time. Controlled by --time-style */
 165 static char const *time_format = NULL;
 166
 167 /* The units to use when printing sizes.  */
 168 static uintmax_t output_block_size;
 169
 170 /* File name patterns to exclude.  */
 171 static struct exclude *exclude;
 172
 173 /* Grand total size of all args, in bytes. Also latest modified date. */
 174 static struct duinfo tot_dui;
 175
 176 #define IS_DIR_TYPE(Type)       \
 177   ((Type) == FTS_DP             \
 178    || (Type) == FTS_DNR)
 179
 180 /* For long options that have no equivalent short option, use a
 181    non-character as a pseudo short option, starting with CHAR_MAX + 1.  */
 182 enum
 183 {
 184   APPARENT_SIZE_OPTION = CHAR_MAX + 1,
 185   EXCLUDE_OPTION,
 186   FILES0_FROM_OPTION,
 187   HUMAN_SI_OPTION,
 188   FTS_DEBUG,
 189   TIME_OPTION,
 190   TIME_STYLE_OPTION
 191 };
 192
 193 static struct option const long_options[] =
 194 {
 195   {"all", no_argument, NULL, 'a'},
 196   {"apparent-size", no_argument, NULL, APPARENT_SIZE_OPTION},
 197   {"block-size", required_argument, NULL, 'B'},
 198   {"bytes", no_argument, NULL, 'b'},
 199   {"count-links", no_argument, NULL, 'l'},
 200   /* {"-debug", no_argument, NULL, FTS_DEBUG}, */
 201   {"dereference", no_argument, NULL, 'L'},
 202   {"dereference-args", no_argument, NULL, 'D'},
 203   {"exclude", required_argument, NULL, EXCLUDE_OPTION},
 204   {"exclude-from", required_argument, NULL, 'X'},
 205   {"files0-from", required_argument, NULL, FILES0_FROM_OPTION},
 206   {"human-readable", no_argument, NULL, 'h'},
 207   {"si", no_argument, NULL, HUMAN_SI_OPTION},
 208   {"max-depth", required_argument, NULL, 'd'},
 209   {"null", no_argument, NULL, '0'},
 210   {"no-dereference", no_argument, NULL, 'P'},
 211   {"one-file-system", no_argument, NULL, 'x'},
 212   {"separate-dirs", no_argument, NULL, 'S'},
 213   {"summarize", no_argument, NULL, 's'},
 214   {"total", no_argument, NULL, 'c'},
 215   {"time", optional_argument, NULL, TIME_OPTION},
 216   {"time-style", required_argument, NULL, TIME_STYLE_OPTION},
 217   {GETOPT_HELP_OPTION_DECL},
 218   {GETOPT_VERSION_OPTION_DECL},
 219   {NULL, 0, NULL, 0}
 220 };
 221
 222 static char const *const time_args[] =
 223 {
 224   "atime", "access", "use", "ctime", "status", NULL
 225 };
 226 static enum time_type const time_types[] =
 227 {
 228   time_atime, time_atime, time_atime, time_ctime, time_ctime
 229 };
 230 ARGMATCH_VERIFY (time_args, time_types);
 231
 232 /* `full-iso' uses full ISO-style dates and times.  `long-iso' uses longer
 233    ISO-style time stamps, though shorter than `full-iso'.  `iso' uses shorter
 234    ISO-style time stamps.  */
 235 enum time_style
 236   {
 237     full_iso_time_style,       /* --time-style=full-iso */
 238     long_iso_time_style,       /* --time-style=long-iso */
 239     iso_time_style             /* --time-style=iso */
 240   };
 241
 242 static char const *const time_style_args[] =
 243 {
 244   "full-iso", "long-iso", "iso", NULL
 245 };
 246 static enum time_style const time_style_types[] =
 247 {
 248   full_iso_time_style, long_iso_time_style, iso_time_style
 249 };
 250 ARGMATCH_VERIFY (time_style_args, time_style_types);
 251
 252 void
 253 usage (int status)
 254 {
 255   if (status != EXIT_SUCCESS)
 256     fprintf (stderr, _("Try `%s --help' for more information.\n"),
 257              program_name);
 258   else
 259     {
 260       printf (_("\
 261 Usage: %s [OPTION]... [FILE]...\n\
 262   or:  %s [OPTION]... --files0-from=F\n\
 263 "), program_name, program_name);
 264       fputs (_("\
 265 Summarize disk usage of each FILE, recursively for directories.\n\
 266 \n\
 267 "), stdout);
 268       fputs (_("\
 269 Mandatory arguments to long options are mandatory for short options too.\n\
 270 "), stdout);
 271       fputs (_("\
 272   -a, --all             write counts for all files, not just directories\n\
 273       --apparent-size   print apparent sizes, rather than disk usage; although\
 274 \n\
 275                           the apparent size is usually smaller, it may be\n\
 276                           larger due to holes in (`sparse') files, internal\n\
 277                           fragmentation, indirect blocks, and the like\n\
 278 "), stdout);
 279       fputs (_("\
 280   -B, --block-size=SIZE  scale sizes by SIZE before printing them.  E.g.,\n\
 281                            `-BM' prints sizes in units of 1,048,576 bytes.\n\
 282                            See SIZE format below.\n\
 283   -b, --bytes           equivalent to `--apparent-size --block-size=1'\n\
 284   -c, --total           produce a grand total\n\
 285   -D, --dereference-args  dereference only symlinks that are listed on the\n\
 286                           command line\n\
 287 "), stdout);
 288       fputs (_("\
 289       --files0-from=F   summarize disk usage of the NUL-terminated file\n\
 290                           names specified in file F;\n\
 291                           If F is - then read names from standard input\n\
 292   -H                    equivalent to --dereference-args (-D)\n\
 293   -h, --human-readable  print sizes in human readable format (e.g., 1K 234M 2G)\
 294 \n\
 295       --si              like -h, but use powers of 1000 not 1024\n\
 296 "), stdout);
 297       fputs (_("\
 298   -k                    like --block-size=1K\n\
 299   -l, --count-links     count sizes many times if hard linked\n\
 300   -m                    like --block-size=1M\n\
 301 "), stdout);
 302       fputs (_("\
 303   -L, --dereference     dereference all symbolic links\n\
 304   -P, --no-dereference  don't follow any symbolic links (this is the default)\n\
 305   -0, --null            end each output line with 0 byte rather than newline\n\
 306   -S, --separate-dirs   do not include size of subdirectories\n\
 307   -s, --summarize       display only a total for each argument\n\
 308 "), stdout);
 309       fputs (_("\
 310   -x, --one-file-system    skip directories on different file systems\n\
 311   -X, --exclude-from=FILE  exclude files that match any pattern in FILE\n\
 312       --exclude=PATTERN    exclude files that match PATTERN\n\
 313   -d, --max-depth=N     print the total for a directory (or file, with --all)\n\
 314                           only if it is N or fewer levels below the command\n\
 315                           line argument;  --max-depth=0 is the same as\n\
 316                           --summarize\n\
 317 "), stdout);
 318       fputs (_("\
 319       --time            show time of the last modification of any file in the\n\
 320                           directory, or any of its subdirectories\n\
 321       --time=WORD       show time as WORD instead of modification time:\n\
 322                           atime, access, use, ctime or status\n\
 323       --time-style=STYLE  show times using style STYLE:\n\
 324                           full-iso, long-iso, iso, +FORMAT\n\
 325                           FORMAT is interpreted like `date'\n\
 326 "), stdout);
 327       fputs (HELP_OPTION_DESCRIPTION, stdout);
 328       fputs (VERSION_OPTION_DESCRIPTION, stdout);
 329       emit_blocksize_note ("DU");
 330       emit_size_note ();
 331       emit_ancillary_info ();
 332     }
 333   exit (status);
 334 }
 335
 336 /* Try to insert the INO/DEV pair into the global table, HTAB.
 337    Return true if the pair is successfully inserted,
 338    false if the pair is already in the table.  */
 339 static bool
 340 hash_ins (ino_t ino, dev_t dev)
 341 {
 342   int inserted = di_set_insert (di_set, dev, ino);
 343   if (inserted < 0)
 344     xalloc_die ();
 345   return inserted;
 346 }
 347
 348 /* FIXME: this code is nearly identical to code in date.c  */
 349 /* Display the date and time in WHEN according to the format specified
 350    in FORMAT.  */
 351
 352 static void
 353 show_date (const char *format, struct timespec when)
 354 {
 355   struct tm *tm = localtime (&when.tv_sec);
 356   if (! tm)
 357     {
 358       char buf[INT_BUFSIZE_BOUND (intmax_t)];
 359       char *when_str = timetostr (when.tv_sec, buf);
 360       error (0, 0, _("time %s is out of range"), when_str);
 361       fputs (when_str, stdout);
 362       return;
 363     }
 364
 365   fprintftime (stdout, format, tm, 0, when.tv_nsec);
 366 }
 367
 368 /* Print N_BYTES.  Convert it to a readable value before printing.  */
 369
 370 static void
 371 print_only_size (uintmax_t n_bytes)
 372 {
 373   char buf[LONGEST_HUMAN_READABLE + 1];
 374   fputs (human_readable (n_bytes, buf, human_output_opts,
 375                          1, output_block_size), stdout);
 376 }
 377
 378 /* Print size (and optionally time) indicated by *PDUI, followed by STRING.  */
 379
 380 static void
 381 print_size (const struct duinfo *pdui, const char *string)
 382 {
 383   print_only_size (pdui->size);
 384   if (opt_time)
 385     {
 386       putchar ('\t');
 387       show_date (time_format, pdui->tmax);
 388     }
 389   printf ("\t%s%c", string, opt_nul_terminate_output ? '\0' : '\n');
 390   fflush (stdout);
 391 }
 392
 393 /* This function is called once for every file system object that fts
 394    encounters.  fts does a depth-first traversal.  This function knows
 395    that and accumulates per-directory totals based on changes in
 396    the depth of the current entry.  It returns true on success.  */
 397
 398 static bool
 399 process_file (FTS *fts, FTSENT *ent)
 400 {
 401   bool ok = true;
 402   struct duinfo dui;
 403   struct duinfo dui_to_print;
 404   size_t level;
 405   static size_t n_alloc;
 406   /* First element of the structure contains:
 407      The sum of the st_size values of all entries in the single directory
 408      at the corresponding level.  Although this does include the st_size
 409      corresponding to each subdirectory, it does not include the size of
 410      any file in a subdirectory. Also corresponding last modified date.
 411      Second element of the structure contains:
 412      The sum of the sizes of all entries in the hierarchy at or below the
 413      directory at the specified level.  */
 414   static struct dulevel *dulvl;
 415
 416   const char *file = ent->fts_path;
 417   const struct stat *sb = ent->fts_statp;
 418   int info = ent->fts_info;
 419
 420   if (info == FTS_DNR)
 421     {
 422       /* An error occurred, but the size is known, so count it.  */
 423       error (0, ent->fts_errno, _("cannot read directory %s"), quote (file));
 424       ok = false;
 425     }
 426   else if (info != FTS_DP)
 427     {
 428       bool excluded = excluded_file_name (exclude, file);
 429       if (! excluded)
 430         {
 431           /* Make the stat buffer *SB valid, or fail noisily.  */
 432
 433           if (info == FTS_NSOK)
 434             {
 435               fts_set (fts, ent, FTS_AGAIN);
 436               FTSENT const *e = fts_read (fts);
 437               assert (e == ent);
 438               info = ent->fts_info;
 439             }
 440
 441           if (info == FTS_NS || info == FTS_SLNONE)
 442             {
 443               error (0, ent->fts_errno, _("cannot access %s"), quote (file));
 444               return false;
 445             }
 446         }
 447
 448       if (excluded
 449           || (! opt_count_all
 450               && (hash_all || (! S_ISDIR (sb->st_mode) && 1 < sb->st_nlink))
 451               && ! hash_ins (sb->st_ino, sb->st_dev)))
 452         {
 453           /* If ignoring a directory in preorder, skip its children.
 454              Ignore the next fts_read output too, as it's a postorder
 455              visit to the same directory.  */
 456           if (info == FTS_D)
 457             {
 458               fts_set (fts, ent, FTS_SKIP);
 459               FTSENT const *e = fts_read (fts);
 460               assert (e == ent);
 461             }
 462
 463           return true;
 464         }
 465
 466       switch (info)
 467         {
 468         case FTS_D:
 469           return true;
 470
 471         case FTS_ERR:
 472           /* An error occurred, but the size is known, so count it.  */
 473           error (0, ent->fts_errno, "%s", quote (file));
 474           ok = false;
 475           break;
 476
 477         case FTS_DC:
 478           if (cycle_warning_required (fts, ent))
 479             {
 480               emit_cycle_warning (file);
 481               return false;
 482             }
 483           return true;
 484         }
 485     }
 486
 487   duinfo_set (&dui,
 488               (apparent_size
 489                ? sb->st_size
 490                : (uintmax_t) ST_NBLOCKS (*sb) * ST_NBLOCKSIZE),
 491               (time_type == time_mtime ? get_stat_mtime (sb)
 492                : time_type == time_atime ? get_stat_atime (sb)
 493                : get_stat_ctime (sb)));
 494
 495   level = ent->fts_level;
 496   dui_to_print = dui;
 497
 498   if (n_alloc == 0)
 499     {
 500       n_alloc = level + 10;
 501       dulvl = xcalloc (n_alloc, sizeof *dulvl);
 502     }
 503   else
 504     {
 505       if (level == prev_level)
 506         {
 507           /* This is usually the most common case.  Do nothing.  */
 508         }
 509       else if (level > prev_level)
 510         {
 511           /* Descending the hierarchy.
 512              Clear the accumulators for *all* levels between prev_level
 513              and the current one.  The depth may change dramatically,
 514              e.g., from 1 to 10.  */
 515           size_t i;
 516
 517           if (n_alloc <= level)
 518             {
 519               dulvl = xnrealloc (dulvl, level, 2 * sizeof *dulvl);
 520               n_alloc = level * 2;
 521             }
 522
 523           for (i = prev_level + 1; i <= level; i++)
 524             {
 525               duinfo_init (&dulvl[i].ent);
 526               duinfo_init (&dulvl[i].subdir);
 527             }
 528         }
 529       else /* level < prev_level */
 530         {
 531           /* Ascending the hierarchy.
 532              Process a directory only after all entries in that
 533              directory have been processed.  When the depth decreases,
 534              propagate sums from the children (prev_level) to the parent.
 535              Here, the current level is always one smaller than the
 536              previous one.  */
 537           assert (level == prev_level - 1);
 538           duinfo_add (&dui_to_print, &dulvl[prev_level].ent);
 539           if (!opt_separate_dirs)
 540             duinfo_add (&dui_to_print, &dulvl[prev_level].subdir);
 541           duinfo_add (&dulvl[level].subdir, &dulvl[prev_level].ent);
 542           duinfo_add (&dulvl[level].subdir, &dulvl[prev_level].subdir);
 543         }
 544     }
 545
 546   prev_level = level;
 547
 548   /* Let the size of a directory entry contribute to the total for the
 549      containing directory, unless --separate-dirs (-S) is specified.  */
 550   if (! (opt_separate_dirs && IS_DIR_TYPE (info)))
 551     duinfo_add (&dulvl[level].ent, &dui);
 552
 553   /* Even if this directory is unreadable or we can't chdir into it,
 554      do let its size contribute to the total. */
 555   duinfo_add (&tot_dui, &dui);
 556
 557   if ((IS_DIR_TYPE (info) && level <= max_depth)
 558       || ((opt_all && level <= max_depth) || level == 0))
 559     print_size (&dui_to_print, file);
 560
 561   return ok;
 562 }
 563
 564 /* Recursively print the sizes of the directories (and, if selected, files)
 565    named in FILES, the last entry of which is NULL.
 566    BIT_FLAGS controls how fts works.
 567    Return true if successful.  */
 568
 569 static bool
 570 du_files (char **files, int bit_flags)
 571 {
 572   bool ok = true;
 573
 574   if (*files)
 575     {
 576       FTS *fts = xfts_open (files, bit_flags, NULL);
 577
 578       while (1)
 579         {
 580           FTSENT *ent;
 581
 582           ent = fts_read (fts);
 583           if (ent == NULL)
 584             {
 585               if (errno != 0)
 586                 {
 587                   error (0, errno, _("fts_read failed: %s"),
 588                          quotearg_colon (fts->fts_path));
 589                   ok = false;
 590                 }
 591
 592               /* When exiting this loop early, be careful to reset the
 593                  global, prev_level, used in process_file.  Otherwise, its
 594                  (level == prev_level - 1) assertion could fail.  */
 595               prev_level = 0;
 596               break;
 597             }
 598           FTS_CROSS_CHECK (fts);
 599
 600           ok &= process_file (fts, ent);
 601         }
 602
 603       if (fts_close (fts) != 0)
 604         {
 605           error (0, errno, _("fts_close failed"));
 606           ok = false;
 607         }
 608     }
 609
 610   return ok;
 611 }
 612
 613 int
 614 main (int argc, char **argv)
 615 {
 616   char *cwd_only[2];
 617   bool max_depth_specified = false;
 618   bool ok = true;
 619   char *files_from = NULL;
 620
 621   /* Bit flags that control how fts works.  */
 622   int bit_flags = FTS_NOSTAT;
 623
 624   /* Select one of the three FTS_ options that control if/when
 625      to follow a symlink.  */
 626   int symlink_deref_bits = FTS_PHYSICAL;
 627
 628   /* If true, display only a total for each argument. */
 629   bool opt_summarize_only = false;
 630
 631   cwd_only[0] = bad_cast (".");
 632   cwd_only[1] = NULL;
 633
 634   initialize_main (&argc, &argv);
 635   set_program_name (argv[0]);
 636   setlocale (LC_ALL, "");
 637   bindtextdomain (PACKAGE, LOCALEDIR);
 638   textdomain (PACKAGE);
 639
 640   atexit (close_stdout);
 641
 642   exclude = new_exclude ();
 643
 644   human_options (getenv ("DU_BLOCK_SIZE"),
 645                  &human_output_opts, &output_block_size);
 646
 647   while (true)
 648     {
 649       int oi = -1;
 650       int c = getopt_long (argc, argv, "0abd:chHklmsxB:DLPSX:",
 651                            long_options, &oi);
 652       if (c == -1)
 653         break;
 654
 655       switch (c)
 656         {
 657 #if DU_DEBUG
 658         case FTS_DEBUG:
 659           fts_debug = true;
 660           break;
 661 #endif
 662
 663         case '0':
 664           opt_nul_terminate_output = true;
 665           break;
 666
 667         case 'a':
 668           opt_all = true;
 669           break;
 670
 671         case APPARENT_SIZE_OPTION:
 672           apparent_size = true;
 673           break;
 674
 675         case 'b':
 676           apparent_size = true;
 677           human_output_opts = 0;
 678           output_block_size = 1;
 679           break;
 680
 681         case 'c':
 682           print_grand_total = true;
 683           break;
 684
 685         case 'h':
 686           human_output_opts = human_autoscale | human_SI | human_base_1024;
 687           output_block_size = 1;
 688           break;
 689
 690         case HUMAN_SI_OPTION:
 691           human_output_opts = human_autoscale | human_SI;
 692           output_block_size = 1;
 693           break;
 694
 695         case 'k':
 696           human_output_opts = 0;
 697           output_block_size = 1024;
 698           break;
 699
 700         case 'd':               /* --max-depth=N */
 701           {
 702             unsigned long int tmp_ulong;
 703             if (xstrtoul (optarg, NULL, 0, &tmp_ulong, NULL) == LONGINT_OK
 704                 && tmp_ulong <= SIZE_MAX)
 705               {
 706                 max_depth_specified = true;
 707                 max_depth = tmp_ulong;
 708               }
 709             else
 710               {
 711                 error (0, 0, _("invalid maximum depth %s"),
 712                        quote (optarg));
 713                 ok = false;
 714               }
 715           }
 716           break;
 717
 718         case 'm':
 719           human_output_opts = 0;
 720           output_block_size = 1024 * 1024;
 721           break;
 722
 723         case 'l':
 724           opt_count_all = true;
 725           break;
 726
 727         case 's':
 728           opt_summarize_only = true;
 729           break;
 730
 731         case 'x':
 732           bit_flags |= FTS_XDEV;
 733           break;
 734
 735         case 'B':
 736           {
 737             enum strtol_error e = human_options (optarg, &human_output_opts,
 738                                                  &output_block_size);
 739             if (e != LONGINT_OK)
 740               xstrtol_fatal (e, oi, c, long_options, optarg);
 741           }
 742           break;
 743
 744         case 'H':  /* NOTE: before 2008-12, -H was equivalent to --si.  */
 745         case 'D':
 746           symlink_deref_bits = FTS_COMFOLLOW | FTS_PHYSICAL;
 747           break;
 748
 749         case 'L': /* --dereference */
 750           symlink_deref_bits = FTS_LOGICAL;
 751           break;
 752
 753         case 'P': /* --no-dereference */
 754           symlink_deref_bits = FTS_PHYSICAL;
 755           break;
 756
 757         case 'S':
 758           opt_separate_dirs = true;
 759           break;
 760
 761         case 'X':
 762           if (add_exclude_file (add_exclude, exclude, optarg,
 763                                 EXCLUDE_WILDCARDS, '\n'))
 764             {
 765               error (0, errno, "%s", quotearg_colon (optarg));
 766               ok = false;
 767             }
 768           break;
 769
 770         case FILES0_FROM_OPTION:
 771           files_from = optarg;
 772           break;
 773
 774         case EXCLUDE_OPTION:
 775           add_exclude (exclude, optarg, EXCLUDE_WILDCARDS);
 776           break;
 777
 778         case TIME_OPTION:
 779           opt_time = true;
 780           time_type =
 781             (optarg
 782              ? XARGMATCH ("--time", optarg, time_args, time_types)
 783              : time_mtime);
 784           break;
 785
 786         case TIME_STYLE_OPTION:
 787           time_style = optarg;
 788           break;
 789
 790         case_GETOPT_HELP_CHAR;
 791
 792         case_GETOPT_VERSION_CHAR (PROGRAM_NAME, AUTHORS);
 793
 794         default:
 795           ok = false;
 796         }
 797     }
 798
 799   if (!ok)
 800     usage (EXIT_FAILURE);
 801
 802   if (opt_all && opt_summarize_only)
 803     {
 804       error (0, 0, _("cannot both summarize and show all entries"));
 805       usage (EXIT_FAILURE);
 806     }
 807
 808   if (opt_summarize_only && max_depth_specified && max_depth == 0)
 809     {
 810       error (0, 0,
 811              _("warning: summarizing is the same as using --max-depth=0"));
 812     }
 813
 814   if (opt_summarize_only && max_depth_specified && max_depth != 0)
 815     {
 816       unsigned long int d = max_depth;
 817       error (0, 0, _("warning: summarizing conflicts with --max-depth=%lu"), d);
 818       usage (EXIT_FAILURE);
 819     }
 820
 821   if (opt_summarize_only)
 822     max_depth = 0;
 823
 824   /* Process time style if printing last times.  */
 825   if (opt_time)
 826     {
 827       if (! time_style)
 828         {
 829           time_style = getenv ("TIME_STYLE");
 830
 831           /* Ignore TIMESTYLE="locale", for compatibility with ls.  */
 832           if (! time_style || STREQ (time_style, "locale"))
 833             time_style = "long-iso";
 834           else if (*time_style == '+')
 835             {
 836               /* Ignore anything after a newline, for compatibility
 837                  with ls.  */
 838               char *p = strchr (time_style, '\n');
 839               if (p)
 840                 *p = '\0';
 841             }
 842           else
 843             {
 844               /* Ignore "posix-" prefix, for compatibility with ls.  */
 845               static char const posix_prefix[] = "posix-";
 846               while (strncmp (time_style, posix_prefix, sizeof posix_prefix - 1)
 847                      == 0)
 848                 time_style += sizeof posix_prefix - 1;
 849             }
 850         }
 851
 852       if (*time_style == '+')
 853         time_format = time_style + 1;
 854       else
 855         {
 856           switch (XARGMATCH ("time style", time_style,
 857                              time_style_args, time_style_types))
 858             {
 859             case full_iso_time_style:
 860               time_format = "%Y-%m-%d %H:%M:%S.%N %z";
 861               break;
 862
 863             case long_iso_time_style:
 864               time_format = "%Y-%m-%d %H:%M";
 865               break;
 866
 867             case iso_time_style:
 868               time_format = "%Y-%m-%d";
 869               break;
 870             }
 871         }
 872     }
 873
 874   struct argv_iterator *ai;
 875   if (files_from)
 876     {
 877       /* When using --files0-from=F, you may not specify any files
 878          on the command-line.  */
 879       if (optind < argc)
 880         {
 881           error (0, 0, _("extra operand %s"), quote (argv[optind]));
 882           fprintf (stderr, "%s\n",
 883                    _("file operands cannot be combined with --files0-from"));
 884           usage (EXIT_FAILURE);
 885         }
 886
 887       if (! (STREQ (files_from, "-") || freopen (files_from, "r", stdin)))
 888         error (EXIT_FAILURE, errno, _("cannot open %s for reading"),
 889                quote (files_from));
 890
 891       ai = argv_iter_init_stream (stdin);
 892
 893       /* It's not easy here to count the arguments, so assume the
 894          worst.  */
 895       hash_all = true;
 896     }
 897   else
 898     {
 899       char **files = (optind < argc ? argv + optind : cwd_only);
 900       ai = argv_iter_init_argv (files);
 901
 902       /* Hash all dev,ino pairs if there are multiple arguments, or if
 903          following non-command-line symlinks, because in either case a
 904          file with just one hard link might be seen more than once.  */
 905       hash_all = (optind + 1 < argc || symlink_deref_bits == FTS_LOGICAL);
 906     }
 907
 908   if (!ai)
 909     xalloc_die ();
 910
 911   /* Initialize the set of dev,inode pairs.  */
 912   di_set = di_set_alloc ();
 913   if (!di_set)
 914     xalloc_die ();
 915
 916   /* If not hashing everything, process_file won't find cycles on its
 917      own, so ask fts_read to check for them accurately.  */
 918   if (opt_count_all || ! hash_all)
 919     bit_flags |= FTS_TIGHT_CYCLE_CHECK;
 920
 921   bit_flags |= symlink_deref_bits;
 922   static char *temp_argv[] = { NULL, NULL };
 923
 924   while (true)
 925     {
 926       bool skip_file = false;
 927       enum argv_iter_err ai_err;
 928       char *file_name = argv_iter (ai, &ai_err);
 929       if (!file_name)
 930         {
 931           switch (ai_err)
 932             {
 933             case AI_ERR_EOF:
 934               goto argv_iter_done;
 935             case AI_ERR_READ:
 936               error (0, errno, _("%s: read error"),
 937                      quotearg_colon (files_from));
 938               ok = false;
 939               goto argv_iter_done;
 940             case AI_ERR_MEM:
 941               xalloc_die ();
 942             default:
 943               assert (!"unexpected error code from argv_iter");
 944             }
 945         }
 946       if (files_from && STREQ (files_from, "-") && STREQ (file_name, "-"))
 947         {
 948           /* Give a better diagnostic in an unusual case:
 949              printf - | du --files0-from=- */
 950           error (0, 0, _("when reading file names from stdin, "
 951                          "no file name of %s allowed"),
 952                  quote (file_name));
 953           skip_file = true;
 954         }
 955
 956       /* Report and skip any empty file names before invoking fts.
 957          This works around a glitch in fts, which fails immediately
 958          (without looking at the other file names) when given an empty
 959          file name.  */
 960       if (!file_name[0])
 961         {
 962           /* Diagnose a zero-length file name.  When it's one
 963              among many, knowing the record number may help.
 964              FIXME: currently print the record number only with
 965              --files0-from=FILE.  Maybe do it for argv, too?  */
 966           if (files_from == NULL)
 967             error (0, 0, "%s", _("invalid zero-length file name"));
 968           else
 969             {
 970               /* Using the standard `filename:line-number:' prefix here is
 971                  not totally appropriate, since NUL is the separator, not NL,
 972                  but it might be better than nothing.  */
 973               unsigned long int file_number = argv_iter_n_args (ai);
 974               error (0, 0, "%s:%lu: %s", quotearg_colon (files_from),
 975                      file_number, _("invalid zero-length file name"));
 976             }
 977           skip_file = true;
 978         }
 979
 980       if (skip_file)
 981         ok = false;
 982       else
 983         {
 984           temp_argv[0] = file_name;
 985           ok &= du_files (temp_argv, bit_flags);
 986         }
 987     }
 988  argv_iter_done:
 989
 990   argv_iter_free (ai);
 991   di_set_free (di_set);
 992
 993   if (files_from && (ferror (stdin) || fclose (stdin) != 0) && ok)
 994     error (EXIT_FAILURE, 0, _("error reading %s"), quote (files_from));
 995
 996   if (print_grand_total)
 997     print_size (&tot_dui, _("total"));
 998
 999   exit (ok ? EXIT_SUCCESS : EXIT_FAILURE);
1000 }